Spam Score: What It Actually Measures (and When to Ignore It)
Spam Score is a Moz metric that estimates how closely a website resembles sites Google has penalized or banned, on a scale of 1–100%. It is not a Google metric, it doesn't detect penalties, and a high score doesn't mean a site is spam — it means the site shares traits with sites that were.
That distinction gets lost constantly: site owners panic over a 40% score, and link builders reject good placement sites over a red number in a free checker. This guide covers what Spam Score actually measures, the 27 signals behind it, what the bands mean, why Google doesn't use it, and where it genuinely earns a place in link vetting.
TL;DR:
- Spam Score shows the percentage of sites with similar features that Moz found penalized or banned by Google — it measures resemblance, not wrongdoing.
- It's built on 27 correlated signals: domain-name traits, link patterns, thin content, and missing "legitimate site" features like contact info or a social presence.
- Google doesn't use Spam Score. There is no spam score inside Google's algorithm, and a high number does not predict a penalty for your site.
- Legitimate sites score high all the time — new sites, small sites, and sites without social profiles trip the same flags spam sites do.
- For vetting link prospects, organic traffic and relevance are far stronger signals; Spam Score works as a secondary flag, not a verdict.
What is spam score?
Spam Score is a Moz metric from 1–100% that represents the percentage of sites with similar features to yours that Moz found to be penalized or banned by Google. Read that definition carefully, because the mechanics matter: Moz took a large set of sites it knew Google had penalized or de-indexed, identified the features those sites had in common, and trained a machine-learning model to measure how much any given domain resembles that penalized group.
So a Spam Score of 30% doesn't mean "this site is 30% spam." It means that among sites sharing this site's feature fingerprint, roughly 30% were penalized or banned in Moz's dataset. It's a statement about a cohort, not about the site in front of you.
Moz launched the metric in 2015, originally as a simple count of 0–17 flags, then rebuilt it as the current 1–100% machine-learning model. You'll find the score in Moz's Link Explorer and in the dozens of free "spam score checker" tools that resell Moz's API data. Since Moz rebuilt Domain Authority in 2019, spam-like patterns also drag DA downward. (Weighing DA against Ahrefs' equivalent? See Domain Rating vs Domain Authority.)
The critical design fact: the model is correlation-based. It has no access to Google's penalty records, no way to detect a manual action, and no insight into intent. It pattern-matches surface features — which is both its usefulness and its central weakness.
What are the 27 signals behind spam score?
Moz's model tracks 27 signals that penalized sites disproportionately share. Moz publishes the list, and it breaks into four rough groups:
- Domain-name traits. Spam-correlated TLDs, unusually long or short domain names, numerals in the domain, hyphens, odd character patterns. Spammers register domains in bulk, and bulk registration leaves fingerprints.
- Link-profile patterns. Low diversity in linking domains, unusual ratios of external links, heavy exact-match commercial anchor text — the footprints of link farms and manipulated link graphs. These carry the strongest predictive weight in the model.
- Content signals. Very few pages, thin content, high keyword density, "poison words" associated with spam niches, abnormal title and meta description lengths.
- Missing legitimacy markers. No HTTPS, no favicon, no contact information, no LinkedIn or Facebook presence, no Google Fonts or Tag Manager installed — the absence of things real businesses tend to have.
Notice what that last group implies. A large share of the 27 signals aren't "spam detected" — they're "normal-site features absent." A brand-new site with eight pages, no social profiles, and no analytics stack trips those flags immediately, not because it's spam but because spam sites are also new, small, and bare. Correlation cuts both ways: the model can't tell a fresh legitimate business from a fresh PBN domain, because on these surface features they genuinely look alike.
That's why Spam Score must be read as a prompt for investigation, never as a conviction. The signals are real statistical correlates of penalized sites, but no single flag — not even five — proves anything about the specific site you're looking at.
What is a good spam score?
Moz buckets the 1–100% scale into three bands:
| Spam Score | Moz's label | What it actually tells you | Sensible response |
|---|---|---|---|
| 1–30% | Low | Few flags triggered; the site's fingerprint matches mostly-clean cohorts | Nothing to do |
| 31–60% | Medium | A moderate cluster of flags — could be spam traits, could be a small or new legitimate site | Look at the site manually before judging |
| 61–100% | High | The site strongly resembles the penalized cohort | Investigate seriously; assume nothing yet |
Two honest caveats on those bands.
First, new and small sites score high by default. A legitimate three-month-old business site can sit at 40–60% purely from youth, page count, and missing social signals. If your site is under a year old, the score is close to meaningless.
Second, the score updates slowly. Moz refreshes these signals roughly quarterly, so fixes you make today can take weeks or months to show. A score is a lagging snapshot, not a live reading.
In our vetting work we treat the bands asymmetrically: a low score tells us very little (plenty of polished link-selling operations score under 10%), while a high score tells us exactly where to look manually. That asymmetry is the single most useful thing to understand about the metric.
Is spam score a Google ranking factor?
No. Spam Score is a third-party estimate with no presence inside Google's systems. Google representatives have said repeatedly over the years that Google doesn't use third-party metrics — not Domain Authority, not Domain Rating, and not Spam Score. Google's own spam policies define what it acts against — cloaking, link spam, scaled content abuse, and so on — and no numeric "spam score" appears anywhere in them.
Three practical consequences follow:
- A high Spam Score cannot cause a penalty. Google acts on its own detection systems and manual reviews, not on Moz's model. If Google has actually penalized your site, you'll see a manual action in Search Console — there's no guessing involved.
- Lowering your Spam Score won't improve rankings. The number affects how you and prospective link partners perceive the site, nothing more. Chasing a lower score for its own sake is cosmetic work.
- Disavowing links doesn't change your Spam Score. The disavow tool is a Google mechanism; Moz can't see your disavow file and doesn't factor it in. People routinely upload disavow files hoping to fix a Moz number — it does nothing, and careless disavowing can genuinely hurt. Our guide to toxic backlinks covers when disavowing is actually warranted (short version: almost never without a manual action).
None of this makes the metric useless. It makes it what it is: a research tool for humans, not a lever on Google.
Does a high spam score mean a site is spammy?
Not by itself — and just as importantly, a low score doesn't mean a site is clean. The model fails in both directions, and anyone using it for link vetting needs to understand both failure modes.
False positives (high score, fine site). New sites, niche sites with small footprints, local businesses without social media, academic or hobbyist sites that never installed Google's ecosystem tools — all of these trip legitimacy-marker flags. Established brands occasionally show surprising scores for the same structural reasons. If you rejected every link prospect above 30%, you'd throw away a meaningful share of the genuine, editorially-run small sites that make excellent links.
False negatives (low score, bad site). This is the failure mode that costs link buyers real money. A professionally built link-selling site — clean design, HTTPS, favicon, contact page, active Facebook, Google Fonts installed — passes nearly every surface check the model runs. Guest-post farms and rebuilt expired domains are engineered to look legitimate on exactly these signals, sometimes complete with inflated Domain Rating to match. The sites most dangerous to pay for often carry the cleanest scores.
In our vetting work the pattern is consistent: Spam Score catches lazy spam — auto-generated domains, bare PBN shells — and misses professional spam almost entirely. The one check that professional spam can't fake cheaply is real organic traffic from Google, which is why traffic sits above every third-party score in a serious vetting process.
How do you lower your spam score?
If your own site's score bothers you, the fix list writes itself from the 27 signals — but be clear about why you're doing it. You're tidying perception (for link partners, buyers, or clients who check the number), not improving rankings.
- Add the missing legitimacy markers. HTTPS, favicon, contact page with a real address and phone, privacy policy, LinkedIn and Facebook profiles linked from the site. These are the cheapest flags to clear.
- Fill out thin sections. More real pages with substantive content reduces the thin-content and low-page-count signals — and helps users regardless of any score.
- Fix metadata basics. Normal-length titles and descriptions, no meta-keyword stuffing, proper canonical tags.
- Audit your link profile for patterns you created. Heavy exact-match anchors and links from obvious networks feed the link-profile flags — and unlike the cosmetic items, these can carry real Google risk at scale. Our backlink quality guide walks through the full audit process.
- Then wait. Scores refresh on Moz's index cycle, not yours. Expect weeks to months before changes register.
What not to do: pay for a "spam score reduction service," or disavow links to move the number (it can't). If there's no manual action in Search Console and your traffic is healthy, a mediocre Spam Score is a cosmetic issue on a working site.
Should you use spam score to vet backlink prospects?
Yes — but as the third check, not the first. Here's the vetting order we use on every placement site before a client pays, and where Spam Score actually fits:
- Organic traffic first. Does the site get real, stable Google traffic — ideally 1,000+ organic visits per month with a flat or rising trend? Traffic is Google's own verdict on the site rendered as a number, and it's the hardest signal to fake. (How to check any site's traffic.)
- Relevance second. Is the site topically related to yours, with a real audience that could plausibly care about the linking page?
- Spam Score third — as a tripwire. A score above roughly 30% doesn't kill the deal; it triggers a manual review: check the outbound-link neighborhood, the content quality, and whether the "site" is a shell. If the manual review comes back clean (often it's just a young or small site), the score is overruled.
- Manual review always wins. No metric — Spam Score, DR, DA — substitutes for opening the site and reading it.
The comparison in one table:
| Spam Score vetting | Traffic-based vetting | |
|---|---|---|
| What it measures | Resemblance to penalized sites (27 correlated signals) | Whether Google actually ranks and rewards the site |
| Catches | Lazy spam: bare domains, auto-generated shells | Both lazy and professional spam — devalued sites lose traffic |
| Misses | Professional spam built to look legitimate | Very new legitimate sites without rankings yet |
| Can be faked | Yes — clear the surface flags | Only expensively; sustained organic traffic is costly to fabricate |
| Cost of a false positive | You skip a good small site | You skip a site too new to have proof — usually fine |
| Verdict | Useful secondary flag | Primary filter |
That's the honest hierarchy: traffic and relevance decide, Spam Score flags, a human confirms. It's the standard behind our link building services — every site shown with live data before payment — and it's why the authority backlinks we deliver come from sites that pass all four steps, not just one gauge.
The honest summary
- Spam Score measures resemblance to penalized sites across 27 correlated signals — resemblance, not guilt
- 1–30% is low, 31–60% medium, 61–100% high; new and small sites score high by default
- Google doesn't use it, it can't cause a penalty, and disavowing doesn't change it
- Low scores don't clear a site: professional link-selling operations pass the surface checks easily
- Vet with traffic and relevance first; use Spam Score as a tripwire for manual review
FAQ
What is a good spam score?
Under 30% is Moz's "low" band and generally unremarkable. But context matters more than the number: a legitimate new site can sit at 40–60% purely from youth and missing social signals, while a polished link-selling site can score under 10%. Read the band, then look at the site itself.
Does spam score affect Google rankings?
No. Spam Score is a Moz metric with no presence in Google's algorithm, and Google representatives have repeatedly said they don't use third-party metrics. Google acts on its own spam detection and manual reviews. A high score can't trigger a penalty, and lowering it won't improve rankings.
Why is my new website's spam score high?
Because many of the 27 signals measure the absence of established-site features: few pages, no social profiles, no contact details, little link diversity. New legitimate sites and new spam sites look identical on those surface traits. Build out the site normally and the score typically falls within a few index updates.
Does disavowing links lower spam score?
No. The disavow tool is a Google mechanism, and Moz has no visibility into your disavow file — the score won't move. Disavowing is only appropriate for manual actions or large-scale scheme cleanup; uploading a disavow file to fix a Moz number does nothing and risks cutting links that help you.
Should I avoid backlinks from sites with a high spam score?
Don't auto-reject — investigate. A high score on a site with real organic traffic, relevant content, and a clean outbound-link profile is usually a false positive. A high score on a site with no traffic and a wall of casino links is confirmation. The score decides where you look, not what you conclude.
Want links vetted on traffic and relevance — not just a gauge? See our pricing: every placement site shown with live Ahrefs data before you pay.